This is the repository for the paper "A Variational Approach to Unique Determinedness in Pure-state Tomography". arxiv-link

🚀 Exciting News! We've launched the numqi package github/numqi, combining all the functionalities of this repository and even more! 🌟 To dive into these features, just install numqi using pip install numqi, and explore the relevant functions within the numqi.unique_determine module. 🛠️

Currently, this repo provides the following functions

1. determining whether a given set of measurement is UDA/UDP or not
2. searching for the optimal measurement for Pauli group, see draft_uda_udp.py/demo_search_UD_in_pauli_group()
3. the UDA/UDP minimum set over Pauli measurements
   • data/pauli-indexB-core.json
   • pauli-indexB-full.json: google-drive-link (around 200 MB)
4. code to reproduce the figure/table in the paper draft_paperfig.py

conda create -y -n cuda118
conda install -y -n cuda118 -c conda-forge pytorch ipython pytest matplotlib scipy tqdm cvxpy
conda activate cuda118

quickstart

from draft_uda_udp import demo_pauli_loss_function
demo_pauli_loss_function() #takes around several hours

# search UD in Pauli groups, default parameters are for 3-qubits, it takes several minutes for one cpu core to run one search
from draft_uda_udp import demo_search_UD_in_pauli_group
demo_search_UD_in_pauli_group()

Every function with prefix demo_ should be runable. The functions in draft_paperfig.py are to generate figures used in the paper.